Ahmed's Integral
Ahmed's Integral
Ahmed's Integral
Zafar Ahmed
Nuclear Physics Division, Bhabha Atomic Research Centre, Mumbai 400 085, India
[email protected], [email protected]
Abstract
In 2001-2002, I happened to have proposed a new definite integral in the American Mathematical
Monthly (AMM), which later came to be known in my name (Ahmed). In the meantime, this
integral has been mentioned in mathematical encyclopedias and dictionaries and further it has also
been cited and discussed in several books and journals. In particular, a google search with the key
word Ahmeds Integral throws up more than 60 listings. Here I present the maiden solution for
this integral.
Z 1
tan1 2 + x2
dx
2
2
0 (1 + x ) 2 + x
(1)
was published [1] in 2001, when it was thrown open to be solved within next 6 months.
Subsequently, 20 authors and two problem solving groups proposed correct solutions. The
solutions to (1) by two authors, Kunt Dale (Norway) and George L. Lamb Jr. (Arizona)
have been published titled Definitely an Integral [2]. AMM usually prefers to publish the
solutions of other solvers than that of the proposer.
This integral is now known as Ahmeds Integral. A Google search Ahmeds Integral
brings more than 60 listings to view. It has been included in encyclopedias and dictionaries.
Various solutions, extensions, properties and connections of this integral have been discussed
in a variety of ways. This integral is very well discussed in two very interesting books on
integrals [3,4] and mentioned in another [5]. This analytically solvable integral also serves as
a test model for various new methods of high precision numerical (quadratures) integrations
[6].
Since in recent times this integral has evoked a considerable attention, I propose to present
the maiden solution that was sent along with the proposal [1].
Let us call the integral (1) as I and use tan1 z = 2 tan1 1z to split I as I = I1 I2 .
Using the substitution x = tan , we can write
Z
/4 cos d
p
I1 =
,
(2)
2 0
2 sin2
2
which can be evaluated as I1 = 12 by using the substitution sin = 2 sin . Next we use
the representation
Z 1
dx
1
1 1
tan
=
(a 6= 0)
(3)
2
2
a
a
0 x +a
to express
Z 1Z 1
dx dy
.
(4)
I2 =
2
2
2
0
0 (1 + x )(2 + x + y )
Further I2 can be re-written as
Z 1Z 1
1
1
1
dx dy
(5)
I2 =
2
(1 + x2 ) (2 + x2 + y 2)
0
0 (1 + y )
=
1Z 1
0
dx dy
(1 + y 2 )(1 + x2 )
1Z 1
0
(1 +
dx dy
.
+ x2 + y 2 )
y 2 )(2
(6)
Utilizing the symmetry of the integrands and the domains for x and y, the second integral
in (6) equals I2 itself. This leads to
Z 1
2
Z 1Z 1
2
dx dy
dx
2I2 =
=
=
.
(7)
2
2
2
16
0
0 (1 + y )(1 + x )
0 1+x
Eventually, we get
I=
5 2
2 2
=
.
12 32
96
2
(8)
References: